ImageReward icon indicating copy to clipboard operation
ImageReward copied to clipboard
verified publisher icon THUDM

w/o prertain loss in the code

Open dongbo811 opened this issue 1 year ago • 1 comments

Formula 3 is in the paper, but not in the code. If I missed it, please let me know.

dongbo811 avatar Jul 23 '24 08:07 dongbo811

We have simplified the code to better demonstrate the ReFL algorithm. You can add it directly inside the code. 😊

xujz18 avatar Aug 01 '24 07:08 xujz18